ML20197C105

Affidavit of J Doyle,Addressing Applicability of Statistical Sampling to Facility

ML20197C105
Person / Time
Site:	Comanche Peak
Issue date:	04/26/1986
From:	Doyle J Citizens Association for Sound Energy
To:
Shared Package
ML20197C088	List: ML20197C079 ML20197C105 ML20197C130
References
OL, NUDOCS 8605130233
Download: ML20197C105 (31)

Text

-

m y-1 iW s

> $3. .DI.!JP, ,

,,3 C 9 .-

UNITED STATES OF AMERICA C Ip g NUCLEAR REGULATORY COMMISSION  %

BEFORE THE ATOMIC SAFETY AND LICENSING BOARD \-

i 9%g M* ^/t h/

s In the Matter of l Docket Nos. 5 445 f;i l and 50- '1 TEXAS UTILITIES ELECTRIC i COMPANY, et al.

(Application for an (Comanche Peak Steam Electric l Operating License)

Station, Units 1 and 2) l AFFIDAVIT OF CASE WITNESS JACK D0YLE The purpose of this affidavit is to address the applicability of statistical sampling to CPSES, based primarily on what has been learned through my review and analysis of pipe supports at Comanche Peak but with clear implications for the entire statistical sampling program proposed by Applicants for CPSES by the Comanche Peak Response Team (CPRT), Stone &

Webster, etc.

THE QUESTION The Applicants would apparently have us address only one issues is the methodology as indicated in the March 28, 1986,-Ropes and Gray transmittal an appropriate procedure? The question more accurately is not whether or not the procedure is correct, but rather is the application proper?

In reference to the application of statistical probabilities to the search for errors in design and construction at CPSES, one must bear in mind what the underlying purpose and intent of statistical probab111tles are. For example 8605130233 860506 5 PDR ADOCK 0500 g 1

O )

1. To establish a level of confidence in the postulated probab111tles:

for instance, a 95% level of confidence indicating that in 19 out of 20 chances the probability is as determined or less, and one chance in 20 that it is greater. In short , the probability would indicate that flaws exist whether or not a single flaw is found in the sample. Beyond this, the level of confidence is 1 out of 20 that the problem is more extensive than assumed.

2. To determine the defects within a sample of significant size in order to project the probability of defects within a significantly larger populat ion.

3. To improve receiving inspection, maintenance cycling, and safety; this is most often the purpose of such techniques (Inspecting a portion of the population to estimate the probable quality to be anticipated in each such population).

(a) In the case of receiving inspection, considerations which govern the sample size and the confidence level are purely monetary; that is, what risk am I willing to assume to avoid the cost of inspecting each article of a shipment?

(b) In the area of maintenance, statistical sampling can serve to establish the time table for preventative maintenance.

(c) Safety can be improved by determining the time f rame for specific inspections or the discovery of potential trouble areas for safety-critical items before the problems occur.

4. Trending to determine the probability of occurrence of the unaccounted for or unpredictable problem (for example, the 0-rings in NASA's Challenger seals) is another area open to statistical analysis.

2

F 9 %

in all of the applications listed above, the paramount purpose is to solve problems -- not to prove that problems are acceptable or to discount their significance. The art of statistical sampling is-utilized to' balance the risk vs. the consequence of a required action, by establishing a time frame for action or a sample size for inspection.

Not to dwell on sampling procedures, but a few of the pitfalls of the ,,

art of statistical probabilities which are applied to design, engineering and construction are worth noting. The principle factors are: the bias of the sample and/or interpretation. Remember, there is an old cliche: There are liars, damn liars, and statistics.

1. Assume, as a hypothetical example, that the NRC wished to determine the safety of diesel generating systems in the hundred or so plants now on line and the principle item of interest was the valve between the fu 1 supply and the diesel engine. If the PRC vere to ask two independent contractors to make a statistical analysis of J he t system based on their expertise, the results could be astounding.

For example, consider the following: (a) The first independent analyst found that all of the valves at the f acilities were in compliance with all of the safety requirements as established in their attribute checklists.

Specifically, the valves in each Instance would fall in the open position, thus assuring a fuel supply to the diesel generator, and thereby insuring that the electrical system would operate the plant protection system in the event of a safe shutdown requirement. (b) For analyst number two, it was found that the same check list results arrived at a totally different interpretation of the valve which f ailed in the open position. This second independent analyst, evaluating his information, deternir;ed that a generic safety flaw had been identified, in that the open valve in the event of an accident represented a potential for intensification of a fire emergency in 3

, i-the diesel generator room. Also, in the above example, since two different conclusions may be drawn from the same criterion, an analyst could exercise

] i his own bias relative to which point he wished to accentuate.

2. The general rule for sampling requires an unbiased sampling

. s technique. In some instances, this methodology is improper; and design, '

engineering, and construction are but three examples of such improper application of what might otherwise be valid statistical sampling techniques.

Assume, for example, that ten individuals are involved in the design,

construction, or engineering of a thousand supports. Any sampling technique

, would Indicate the level of confidence in the derived reliability but would i

fall to identify the somewhat knowledgeable individuals within the group.

In the case of engineers, it is imperative that the sampling techniques be

] directed on a biased basis towards tne Individuals involved to insure that 7 the product constructed is the product desired and, if not, which 4

Individuals are responsible, then proceed from this new benchmark in 1

' combination with resampling procedures.

I I

APPLICABILITY OF STATISTICAL SAMPLING TECHNIQUES As shown above, sampling techniques have a valid place in industry, but i

not for the purpose of evading or dismissing problems. In the current i

situation with CPSES, the sampling techniques employed in the reinspection effort would serve no useful function for areas not subjected to a '

corrective program, since in at least two specific areas (pipe supports and cable trays) CASE has already demonstrated that major def ects exist and '

i Applicants are (to say the least) prone to downplay the allegations. Beyond this, Applicants have not demonstrated that they are quallfled to generate a i

- 4

-m - ~ _ _ -,,- - . - . - , , _ __,..-,__m-- . , , - _ , , , . - . . - . - - - . , , _ , . , - . _ _ , - - - - .

,s Y Y j s list of attributes which would satisfy the 95/5 criterion. (See previous disersssions in CASE pleadings, such as CASE's First Critique of Applicants' ,

Comanche Peak Response Team (C RT) Plan, sent to the Board under cover letter of 8/14/85.) '

Thereare'severalfeaturIsofthisprinciple(statisticalsampling) ,

which are not applicable to CPSES or any other like situation:

~

1. . The concept of statistical sampling to qualify a population is l based on an idealized environment. That_is, the sttempt to create a i co'nforming population is the desike of ; and within the capability of, the n

producer. The purchaser also has the desihe and the ability to perforts an unbiased sa.:spling approach. And ultimately it must' be determined whether or not thefapplicability of tha te :hnique is proper.

As a case in point, tl'e aircraf t industry has the desire to produce a fa11 safe product for several reasons: First, to attain a name on the world market for a safe; product; second, to avoid product malfunction within the anticipated lifetime of the aircraf t; but third, and most important, to protect the life of the traveling public. To this end, oversight by the FAA

).

and the aircraft corporation is critical. Without going into the checks and balances system required to achieve an end product which meets these goals, f

it is noteworthy to mention _that type certification is dependent on the

(

final test program. The final test program within the aircraft industry is unique, in that it involves t"ne testing of the first three completed

~

airfranes'to destruct 1on. The purpose of this is to insure that all of the

' assumptions, etc., used to produce the various components (each previously

!. tested Individually) which are intende'd to resist the various flight loads A

(that is, stetic, fstigue, and dphamic loads) have achieved the results m _c

( '. , .

desired and that these components Jill not fall when assembled into the h/ ,

l l 5 l

p,

)C a i . , .;

-r -evJr -4.--w-- m4 __ s,..- ,-- - ,_ , , __ __,, . _ - ....,..m m _,,_ _ _ g _ _ . ;_ , , , _ _ _ , . _ _ , . , , _ , , _. . , _ _ ,

T I final integrated airframe.

Within the aircraf t industry, to insure that there will be no failure at the final test stage, statistical sampling is utilized during the development phase for trending and other purposes. The end product for the aircraft producer must be flawless within the human capab111tles of the producer. To detect flaws in assumptions or discrepancies in design is of more importance than schedule or budget. Following this program is a flight test program, and anyone within this industry who suggested that it would be prudent to replace these costly procedures with a statistical sampling technique would be immediately escorted to the nearest funny farm.

The nuclear industry has no such test programs for structures; therefore, in light of the more significant risk vs. consequence factor, it is imperative that the nuclear Industry do it right before the unit become s operational.

2. A major consideration for statistical sampling involves the parameters considered and the interpretation of the results. In the case of CPSES, both Applicants and the NRC Staff, when told of the specific f aults within Unit 1, confronted the problem with three views. First, the argument was that the allegations were without merit. Second, they argued that if there were any errors, Applicants would catch them in their final vendor certification program. And finally, the NRC Staff did a statistical sample on a group of 100 supports, randomly selected f rom a population of 1,264 vendor certified supports, for the Staff's independent design verification i

in answer to CASE's allegations. The statistical analysis was based on M11

Standard 105D-63, " Sampling Procedures and Tables for Inspection By Attributes," and included only those supports which had completed I

i Applicants' vendor certification program (i.e., the pipe support design organization had reviewed the supports and certified that each of them were

6 l

r l

• i capable of resisting the design loads as determined from the architect / engineer's pipe stress analysis). The results were that , of the fif teen attributes checked, not one defect was found among the 100 supports checked. (See NRC Staff's Special Inspection Team (SIT) Inspection Report 50-445/82-26, 50-446/82-14, NRC Staff Exhibit 207, bound in following Tr.

page 6289, 5/17/83, Vol. II, at pages 2, 3, 6, 7, 9, 55 through 58, especially 55 and 58.) If one were to have accepted the NRC Staff's evaluation of the status of the engineering for the pipe supports at CPSES in February of 1983, the date of the report, the plant would now be on line (although perhaps not operable).

The statistical sample as shown by the NRC exceeded what is now projected as an adequate initial sample to assure reasonable confidence that the plant has been safely designed and constructed. Statistically ^, the NRC Staff's findings would indicate a 97% reliability factor at a 95% level of confidence based on simple binominal distribution. The new program suggested by Applicants is siellar to the program executed by the NRC Staf f with one majot exception. This new attempt to use the statistical sampling methodology would be executed by the Applicants themselves or at least under their control -- hardly an unblased approach.

The intended approach by Applicants is without merit, since there are no avenues to alter reality by such application even if it were done by

proper statistical sampling. And the facts are that Applicants are groping in the dark for any means to lessen the impact of the failure of large numbers of the pipe supports at CPSES and, of more consequence, to avoid uncovering another Pandora's box in what are yet unchallenged areas.

l i

An example of this blind grasping for excuses, justifications, negating l

of necessity for the faulty supports, or the wild use of semantics may be 1

7 I

i i 4

noted in the following single example

. When the allegations of the problem of instability were first Introduced, Applicants and the NRC Staf f both took the position that for i instability to be a problem, there had to be more than one support in a row which was unstable. But as was noted during the hearings, the NRC Staff has problems when inspecting areas to which they are directed. For example, in CASE Exhibit 669B (Attachment to Doyle Deposition / Testimony, accepted at Tr.

3630), Exhibits 4A, 4B, 4S, and 40. I referred to two unstable supports on the main steam system. In 4S and 40, these unstable supports were MS-1-001-t I

005 and MS-1-001-003, which were located on the same main steam line in the Safeguards Building. The NRC Staff later stated that they had looked at each area of concern and each support that were noted in CASE Exhibits 669, 669A, and 669B (Doyle deposition / testimony and attachment, all accepted at Tr. 3630); however, the NRC Staff stated during the hearings that they saw no case with more than one unstable support in a row, even though the above-referenced supports were but two of five unstable supports in a row on that main steam line. The Applicants made no attempt to correct the record, although they were involved at that time and earlier in correcting the cluster of five unstable supports or at least four of them. (See Affidavit of CASE Witness Jack Doyle, CASE's 10/6/84 First Motion for Summary Disposition Regarding Certain Aspects of the Implementation of Applicants' Design and QA/QC for Design, pages 14 and 15.)

l Another excuse offered was that the up and down stream supports would stabilize the faulty support. The problem of instability proved beyond Applicants' ability to dismiss, and the game of semantics couldn't hold up. <

i So Applicants then offered four methods to correct the problem, all of which failed. These corrections, as the Licensing Board will recall, were:

bumpers for some of the unstable supports, lugs for others, cinching up of

! 8

? I the U-bolts for the majority, and clip angles with U-bolts for box frames to prevent frame rotation. (See Cygna's 2/19/85 letter which was attached to CASE's 2/25/85 Notification of New and Significant Information and CASE's Supplement to CASE's 10/15/84 Motions and Answer to Applicants' Motion for Summary Disposition Regarding Stability of Pipe Supports.) And as always, whenever Applicants are lef t with no other avenue but corrective action, they resort to their time-worn, off-the-cuf f statement that the impact had no safety significance, even when the defects number in the tbcusands.

Finally, it seems senseless for Applicants to suggest that statistical

. sampling will prove anything when in fact we have several major sampling efforts from the past (as discussed in more detail in the following) aside from the NRC SIT flop (see above) -- by which it has been shown that many generic defects exist at CPSES.

i First, for example, there was my original deposition and the back-up documentation which involved only some 58 supports, which has over time been proven to be accurate. These documents are the basic source of the 19 generic allegations alluded to by the NRC in the S1T Report. As a per cent of this sample of 58, we have shown that 31% are defective thus far (see j pages 7 through 18 in CASE's 10/6/84 First Motion for Summary Disposition).

l Second, there was the attempt by Applicants to address the stiffness I

! allegation which was danced around for months. Finally, Applicants ordered a stiffness study. While the system and supports were selected by Applicants and the analysis was stated to be proof positive that variations in stiffness had a minimal ef fect on the supports and in no case led to overstressed conditions, the fact is that Applicants noted in the raw data, but f ailed to elaborate on the f act that a significant increase in load

)

9

m a i resulted from random variations in stiffness. With the same raw data generated by Applicants, CASE found for one system containing a single support with an actual stiffness of 1/70 of the generic stif fness, the raw i data revealed the following: 20% of the supports in the system exhibited

, load increases of more than 25%. The greatest load increase was for a

, support which went from 824 lbs, to 1371 lbs. At one anchor the forces and y

one moment increased more than 25%; at the other anchor, all of the moments and two forces increased. The actual facts, therefore, prove that in 'a i statistically significant portion of this sample, the loads used in the original design are non-conservative; and any reference to "no significant I

impact" on the allowable stress ratio is not sound engineering evaluation but pure dumb luck, because properly designed systems are not so conservatively designed that one could enjoy such margin for error. (See 10/14/83 NRC Staf f Witness Dr. Chen's Af fidavit on Open items Relating to

, Walsh/Doyle Concerns, and its attached Applicants' 8/17/83 Additional Pipe Support Generic Stiffness Study.)

The third statistical sample complied by Appil ants involves the tests

. for U-bolts acting as two-way restraints. (See Applicants' 5/23/84 Motion for Summary Disposition, Table 3, Attachment to lott.1/Finneran Af fidavit, on

, U-Bolts Acting As Two-Way Constraints.) From Applicants ' ot'n af fidavit and attachments,1 found the following: 8 suppotts were redesigned; 9 others l had significant load increases over the loads used in the original design; l

and a minimum of one support failed -- 17 deficiencies out of a sample (in l

I fact, according to Applicants, 70 was the total number of such supports).

This is certainly a significant rate of failure -- significant enough so as not to require further sampling to prove that the design is flawed.

Additionally, Applicants have since corrected their own affidavit, admitting that the number of U-bolts which must be replaced has been changed from the 10 l

l t

O e original 8 to 22, and the total number has been changed from 70 to 76, which Indicates a failure rate of 29% of the total, and this is according to Applicants themselves. (See Applicants' changes in affidavits attached to their 1984 Motion for Summary Disposition, 11/85.)

For the fourth sample, we shift to Cygna Energy Services and their Phases 1 and 2 report. Of the 9 supports with calculations, the Cygna review found no problems and that everything was going according to plan.

In my review of the same 9 calculations supplied by Cygna, I found major errors in engineering calculations in 6 of the packages and, in f act , in one case the support failed (as it turned out later, this was a generic problem, as there were several others of this type of f ailure elsewhere, although Applicants never stated anything on this point until af ter we brought out the fact that this configuration failed). Now here we have 67% of the 9 supports containing engineering flaws, and one of the 9 is a failure. This is definitely a significant sample, clearly demonstrating a significant reduction in the anticipated reliability of a nuclear power plant system.

(See CASE's 10/6/84 First Motion for Summary Disposition, page 22.)

The fifth sample is also of Cygna's origin, the Phase 3 study done by Cygna. (See 10/6/84 CASE's First Motion for Summary Disposition, affidavit pages 49, 51, 72, 73, and 78.) Of the area examined by Cygna, their sample contained 22 supports for the main steam system. Within these, I found the following: of the 22 main steam supports,18 had gross errors in engineering and one support required major modification due to a design failure (this information came from Cygna's raw data), but Cygna did not elaborate on the meaning of this data nor the potential consequences thereof during the life of the plant. Here again, we have a significant percentage 11

__. .m . ._ _ _ _ _ _ _ _ _

of design deficiencies which caused no alarm for Cygna, Applicants, or the i NRC Staff.

In view of the above, it seems incredible to assume that Applicants can show by any method of statistical sampling that the various systems of supports at CPSES are acceptable, unless of course they intend to perform i

~

their analysis in the same way that the NRC used in the S1T Report to prove that CPSES was ready for a license in February 1983. Beyond all of this, the use of statistical sampling does not comply with the intent of 10 CFR Part 50, Appendix B, Criterion XVI (as will be discussed in more detail later herein under CONCLUSIONS).

Ih*rERPRETATION OF SAFETY SIGNIFICANCE Applicants are pressing forward in at least one area: they have moved ,

d up from oracle of qualitative engineering to oracle of safety-significance.

For years it has been Applicants' firm and unswerving position that all of the supports at CPSES were properly designed and constructed. Within the recent past, however, Applicants' house of cards has started to crumble.

The new party line is that, while the support does not comply with 1

commitments which were in place at the time of the original design, it is of

, no consequence, since the support in question has no safety-significance.

This line has been used for individual supports, dozens of supports, hundreds of supports, and now thousands of supports (see Applicants' April 18, 1986, SEC Form 8-K filing, copy of which it is our understanding has already been provided to the Board and parties by Applicants).

By altering the plane of argument from the technical to the semantic, Applicants hope to avoid responsibility for their past actions. This results when a generic semantic cloud is introduced (for example, "no 12

- - . ~ - . _ . . _ _ . - . = _ _ - _ . _ . _ . . _- ___ __ _ _ _ - _ . - _ . . ,

m

safety-significance"). When this argument is of fered, CASE is faced with an Orwellian task -- to do a fault tree analysis or SCA (sneak circuit analysis) to prove that singly or collectively the problems have a safety-significant impact on the plant. By their claims of no safety-significance, Applicants have moved the concept of hyperbolic imperatives one step beyond.

As for myself, I find it difficult to offer logical arguments in this world

, of illusions created by Applicants to attempt to avoid technical responsibility.

And beyond this, I don't believe CASE must bear the burden of proof on the " safety-significant" issue on the issue of fault tree analysis. It is Applicants' argument -- let them prove that Congress intended that nuclear i

power plants with thousands of discrepancies (by Applicants' own admission) do not represent safety-significant issues.

i The fact that the burden of proof shouldn't be on CASE is borne out by the ??RC Staff. When asked about the significance of unstable supports during the hearings, Judge Bloch asked NRC Staff Witness Mr. Taylor if the effect of one support failure should be considered, to which Mr. Taylor relled "You are asking for a failure mode analysis for every run." (See Tr.

6999.) When engineering does not exceed the code limits, the safety 4

question is answerable based en historical and existing analytical data;

1. however, when the. codes are not adhered to, we have established a new baseline with no handle on future events related to the deviations.

Relative to the above, I have not noticed that the NRC Staf f has volunteered

! to rake a stand on the " safety-significance" issue and has thus far accepted Applicants' visions of safety significance.

Once one deviates from the accepted standards for engineering and design, the reliability of the facility can only be adversely af fected. The d

13 f

t fact that, for a riven point in time, it can be shown that total f ailure will not result from overload is without relevance. I have said myself that any idiot within the industry knows that you can exceed code allowables by P

no less than a factor of two during a static test without experiencing a failure; this is due to the safety factor utilized for design. Therefore, by offering testing or analysis to show that a portion of the safety factor remains and that such showing would relleve the designer of the responsibility for an overload or exceeding allowables on the grounds that a 20% deviation from the cede requirement would not constitute a flaw resulting in collapse is the height of absurdity, since the fact is that the code requirements have been violated and quality has been compromised over 4

the life of the plant, thereby rendering the plant indet e rminate. It is no longer possible to accurately state what the probable survivability of this plant is over time. The safety margins have been undercut and you can no longer say that you can rely on that factor of safety. The real safety significance for this plant is Applicants' inability to comply with standard codes and practices.

Once the Applicants took the position that it is more expeditious to remove, redesign, reinstall, reinspect supports or to redesign and modify supports than it is to qualify the as-built conditions, Applicants conceded that the support. systems are not engineered but in fact are abstractions, l for if the supports in the s' stem y were engineered to create the arrangements and configurations of the components, the proof of adequacy would precede the design and construction and not be required to be tailored to an

, ambiguous as-built system generated by guesswork. In short, the Applicants violated the fundamental rule of system reliability: " KISS" it (i.e., " keep it simple, stupid"). The result is that, not only have Applicants found it impossible to qualify their abstract supports for the Licensing Board, they i

14

now can't even justify the supports to themselves. (See Applicants' Management Plan, June 28, 1985, page 67; also see corrections to Iotti/Finneran affidavit relative to their motions for summary disposition, and Applicants' April 18, 1986, SEC Form 8-K filing.)

Safety significance cannot be prophesied by great oracles nor negated by decree, particularly when the decree is issued by the perpetrator of the discrepancy requiring explanation. The constant claims of "no problem" by Applicants are in themselves safety significant. The wild unsubstantiated claims of no safety significance Indicate that Applicants are not quallfled to participate in any unblased sampling technique -- because they are as convinced now that there is no safety significance to the problems as they were originzlly convinced that there were no problems in the first Instance.

With upwards of 1/3 of the supports in Unit 1 requiring modification and others pushing the margins of safety to the point where confirmation is required on a pending change in a code, Applicants would have us believe that they and they alone are in the unique position of determining the long-term safety impact merely by prophesying that there is no safety significance. It makes one wonder why such vast sums of time and money are spent on " fault tree analysis" or sneak circuit analysis (SCA) when right down at Comanche Peak we have the prophets who can determine right of f the wall what is acceptable and what is not, and they can do it with a straight face. Among other factors, it was this attempt (by the engineering industry in general) at projecting performance beyond their capacity that led the American Society of Civil Engineers (ASCE) to redefine responsibility; the l ASCE found that the engineering improved between 1924 and the period of the 1960's, whereas in the period of the 1960's engineering practices 'aave been l

i l 15

1 o I a 3 deteriorating (see Attachment A hereto, page 3, 1.2, second paragraph, of ASCE " Final Report and Recommendations on Assignment of Authority &

Responsibility for Design of Steel Structures," October 20, 1985).

In reference to the preceding, it has been the practice over the years for engineering corporations to dismiss certain areas of design as not important to the structural integrity of the facility. And the American Society of Civil Engineers has come down on this cavaller approach to ^A engineering (see Attachment A hereto.) This has only occurred because of a rash of terrible failures to structures which were believed safe: for example, the Hartford sports center, the Kemper auditorium, and the walkway ,

at Kansas City which killed 113.

As far as the nuclear industry is concerned, a few of the major problems with predicting accidents can be noted in the following

" Insignificant" incidents which resulted in major consequences. The first occurred at the Browns Ferry Plant in Alabama, where the facility came to a near-disaster due to a fire which was started by a worker's candle while performing an inspection; he inadvertently set fire to the insulation, which in turn resulted in damage to electrical cables vital to the plant protection system. A safe shutdown of the plant resulted f rom jury-rigging of the electrical system around the damaged cables and further damage to the plant was thereby averted. The insulating material which caused the fire to spread was not considered safety-significant prior to the accident.

The second incident was the Fermi 1 unit located at Monroe, Michigan.

l The accident which destroyed this plant was close to a catastrophe prior to i control being restored. This accident resulted from the blockage of the coolant by a zirconium deflector which not only wasn't considered to be safety-significant, it wasn't even required.

t 16 1

. . _ _ . _ - - - . - - - _ - - ,~ _- _..

. a Third, the ALPR-1 reactor in Idaho suffered a disasterous chemical explosion due to a safety device. A control rod became stuck, and when dislodged by a worker, it withdrew in excess of safe limits on the rod movement. The results were disaster.

Fourth, the support system for the steam generator at Shippingport i failed at loads well below what were assumed to be safe limits, nearly resulting in a disaster for this plant. The only factor which averted a major catastrophe was the fact that the unit was just being filled with water and was not operational at the time.

Finally, there was Three Mlle Island (TMI). It is ironic that, here again, we have a situation involving minor problems which collectively spelled disaster. But first, the disaster at TMI also had a bright side in that it shocked America out of complacency in regards to what may cause accidents and, in fact, introduced us to an accident which was not possible, as was pointed out during the hearings by John Kemeny before his committee formulating a report for the President. Fortunately, this benefit was bestowed at a relatively insignificant cost, mainly a monetary loss. Such lessons most often are acquired by a massive loss of life.

The factors which contributed to the TM1 accident were individually of minor significance, and in fact, each malfunction had been noted as occurring prior to the accidenc, which was caused by their collective interaction. The contributors were: (1) A pilot-operated relief valve failed in the closed position; the indicator lights on the operator's panel were designed to announce the last command issued (not the actual valve position), the operator was unaware of the actual valve position. This simple sequence of events deprived the reactor core of coolant for over two 1

hours. (2) The core temperature thermal sensors were capable of detecting i

I

! 17 l

l i

- - , - - . - . ~ . . - . . ~ . _ _ , _ . - - .

.. o temperatures of the fuel to several thousand degrees; however, since this was an accident that couldn't occur, the readout for the operator was limited to only 700 degrees. Therefore, the initial problem went undetected. (3) The condensate water by pass system, which could have supplied cooling water to the reactor, was operable only on a manual basis and had design flaws ttiat made operation during this accident next to impossible. (4) The operator overrode the computer system which activated the Emergency Core Cooling System (ECCS) and thereby ensured the accident's progression. (5) Later in the accident, the computer output fell several hours behind the events.

The results of these sequences of events outlined above led to the

" Report of The President's Commission on the Accident at Three Mile Island,"

October 1979 (the Kemeny Report). Some of the contributions to this report and some of the findings are worth noting. One item is in reference to the mission of the NRC and others and changes required. It probably addresses the current absurdities being perpetrated at CPSES best. The conclusion to the report contains the following sage advice (page 7, OVERVIEW, OVERALL CONCLUSION):

"To prevent nuclear accidents as serious as Three Mlle Island, fundamental changes will be necessary in the organization, procedures, and practices -- and above all -- in the attitudes of the Nuclear Regulatory Commission and, to the extent that the institutions we investigated are typical, of the nuclear industry." (Underscoring in the original.)

Similar conclusions were drawn in the Rogovin Report ("Three Mile Island, A Report to the Commissioners and to the Public," Nuclear Regulatory Commission Special Inquiry Group, January 1980).

Apparently this message has been nisplaced or displaced in the archives of Applicants' collective mind. I use the term " mind" advisedly, for it is 18

. w

~ difficult to Imagine that anyone actually believes that the intent of .

Congress was to create plants intentionally that were designed and constructed to the 95/5 criterion.

CONCL1'SIONS The concept of statistical probability as a tool to determine the safety status of a plant plagued with problems is outside the ecope of the art. Some of the obvious reasons for this in the case of CPSES are as follows: >

1. Attributes must be known in advance, and Applicants have proven by ,

the manner in which they have dealt with the large-bore pipe support systems that they do not know what potential problems may occur, and even when told by others, they are prone to dismiss them as being of no safety-significance.

Further, Applicants are admittedly utilizing depleted j attribute lists, which still further negates any possibility of their statistical sampling achieving the purpose for which they seek to use it.

! 2. In the case of CPSES, it cannot be argued that the errors leading I

to the contests over the last four years have been isolated. On

i. the contrary, CASE first outlined a massive list of discrepancies 1

l In the pipe support area. CASE's investigation of Cygna's Phases 1 and 2 Report uncovered numerous discrepancies (and led to i Cygna's further more intensive review) of the cable tray design and construction. The NRC's Technical Review Team (TRT), in its j review of hundreds of allegations by whistleblowers and CASE witnesses, uncovered numerous faults with the construction in 19 l

- _ -. _ . . _ . . _ _ . _ _ _ _ . _ _ _ _ _ _ . _. _ _ . _ . _ _ . ~ . . _ .- - - -

. - _ . -_ _. . =. .- - - -

o- .

general -- as well as confirming the allegation (previously dismissed by NRC Region IV as being without basis) that there were problems with the design of the Control Room itself. The NRC's Construction Appraisal Team (CAT) also noted a number of discrepancies bafore them. And the Atomic Safety and Licensing Beard itself discussed many specific problems in its 12/28/83 Memorandum and Order (Quality Assurance for Design); etc., etc.

3. The CPSES bias is a legend in its own time. First, Applicants and NRC Staff's claim has been that there are positively no i

allegations that have merit, followed by the claim that none of the errors are safety-significant. Beyond this, if Applicants' position is that there are no real problems, what motivates them j to alter this position and perform what they claim will be statistical sampling which is not biased in favor of the Applicants?

4. The initiation of a statistical sampling program will not correct 3

! problems which are located in positions which are not revealed by the program.

5. Institution of statistical sampling is contrary to the purpose of engineering, which is to develop structures which comply with

, codes and standards 100% -- not to show that the plant has only a specific level of problems within an odds frame of 19 out of 20 chances.

6. To solve problems, or in fact to determine if problems exist or are significant, one must be able to define the problem and its causes (we agree with the Board on this point; see Board's 4/14/86 Memorandum and Order (Motion to Compel Production of Checklints)).

i Since Applicants do not appreciate this correlation, we do not 20 c-- , --i , w y-- -- --v-- , , ,m --

, _ -, ,,,, ..-, - _ . , , , - w - ,-w o, w-- -- --.a--------.-_-.~--,-e- w s--v--e-e- -

. . . . -. -- . _ - . ~ . - .

believe that Applicants could execute a proper statistical sampling exercise to qualify any system at CPSES.

7. The directors and officers or the utility corporation serve at the pleasure of the stockholders (not the public). The upper management serves at the pleasure of the board and its officers.

And lower and middle management serve at the pleasure of their upper management. The lower and middle management who are involved in the day-to-day execution of such programs as statistical sampling are effectively the same lower and middle ,

management which directed the program which resulted in the vast array of problems outlined above. Now this same group is supposed to institute a program which will show that various aystems are properly designed and constructed. This represents a classic case of conflict of interest.

8. 10 CFR Part 50, Appendix B, Criterion XVI, Corrective Action, states:

" Measures shall be established to assure that conditions adverse to quality, such as f ailures, malf unctions, deficiencies, deviations, defective material and equipment, and nonconformances are promptly identified and corrected. In the case of

! significant conditions adverse to quality, the measures shall assure that the cause of the condition is determined and corrective action taken to preclude repetition. The identification of the e significant condition adverse to quality, the cause of the condition, and the corrective action taken shall be documented and reported to appropriate levels of management." (Emphasis added.)

The words are clear in Criterion XVI: " conditions adverse to quality." There is no requirement that such conditions must first satisfy a determination of safety significance, since all elements 21

. -a of components classified A, B, C, and MC in a nuclear power plant carry safety-significance. On this criterlon, the first sentence p

' relates to conditions less than significant conditicas, which are covered in the second sentence and thereafter. Adverse conditions which are only predicted, but randomly located, cannot be corrected. Therefore, the premise of statistical sampling le

]

l- counter to the requirements of Crlterion XV1.

9. An important feature lacking for producing a significant statistical sample by Applicants is attitude. For that matter, Applicants' attitude is that the principal factor for consideration is " safety-significance," and this, only by Applicants' personal interpretation. Further, they are convinced that a plant with zero defects is not achievable. In fact, Applicants state that Appendix B was instituted with the 4

understanding that errors will exist but will lack significance due to the " defense In depth" philosophy. (See Applicants' 1/31/66 Memorandum in Response to Board's Memorandum (Statistical Inferences from CPRT Sampling), page 15, top of page and Footnote 4.)

Tt.e fact that producing error proof nuclear plants is next to impossible is not arguable. What is arguable is the cavaller attitude that we can accept a 5% flawed system based on a 95%

level of confidence (and CASE does not believe that Applicants can even hope to achieve the 95/5 confidence level). The fact that errors elude diligence does not mean that once error is suspected it can be overlooked on the grounds that one knew it would happen 22

-- ~ _. .-

anyway. Additionally, the term " defense in depth" refers to the complex redundance and overlapping safety systems in place to J

mitigate the potential consequences of an accident, that may result from unknown or unaccounted for conditions leading to a rare event or errors which have eluded or evaded the best ef forts of engineers -- and not to allow for suspected flaws to be dismissed on a 95/5 basis. All references to " reasonable assurance" carry tha implied belief that Applicants must have exercised every effort to ensure that the nuclear power plant has been accurately designed, constructed, and inspected to the limits of human ability to insure the public safety.

10 CFR represents the federal position for implementing the Atomic Energy Act, as amended. (See more detailed discussion in CASE's 2/3/86 Response to Board Memorandum (Statistical Inferences from CPRT Sampling), pages 2 through 8.)

In no sentence in any NRC implementing regulations, and specifically in 10 CFR Part 50, Appendix B, is the thought conveyed that error is acceptable.

Although it may be impossible to totally avoid error in design and construction, certainly it is not proper for an applicant to use this as an excuse for dismissing errors and defects as not being safety-significant. Further, Applicants' position is inconsistent within itself, since their statement in Footnote 4, page 15, of their 1/31/86 pleading (that there will be defects even with a OA/QC Program that is fully compliant with Appendix B) underscores the need that at a minimum the requirements of Appendix B must be followed (rather than the opposite conclusion which Applicants would have the Board adopt).

23 2

y __, , _ _ _ _ , , ~ . , , . _ _ - . . . . _ . . , . . - - _ , . . , _ - - . - - _ - - , _ , - . - - , . - - _ ___ _ _ _ _ , , _ .-

. + +

For the above reasons and those previously articulated, we believe firmly that the art of statistical sampling is not applicable for the Purposes intended at CPSES.

9 24

.. o I have read the foregoing affidavit, which was prepared under my personal direction, and it is true and correct to the best of my knowledge and belief.

ml (Sigg ~~p Date: Alifil 2G / 98G3 STATE OF [(, M SM OSs M COUNTY OF l 1 Om TL t On this, the RC, day of p, r ,198_k personally appeared

<E

t. _ AC JOVL6 a

, known to me to be the person whose name is subscribed to the foregoing instrument, and acknowledged to me that he executed the same for the purposes therein expressed.

44 -

Subscribed and sworn before me on the JV day of B/L ,

198h.

1NA the Notary)

State 6fPub [lic/Jnand _um7T

, Lfhe;5 -

(-

My Comission Expires: [O fo

ATTACHMENT A i AMERICAN SOCIETY OF CIVIL ENGINEERS l

FINAL REPORT AND RECOMMENDATIONS ON ASSIGNMENT OF '

AUTHORITY & RESPONSIBILITY FOR DESIGN OF 1 STEEL STRUCTURES '

i i

! October 20, 1985 ,

I l

i t .

V .

a e o held in Boston. MA. on June 16-17.1983, and reads

1.0 INTRODUCTION

as follows:

"To imestigate the perceived problem of the 1.1 Purpose transfer of cenain aspects of responsibility for The American Society of Civil Engineers design of steel structures to the structural steel believes that to protect the safety of the public, a fabricator and to make recommendations for continuing re>ponsibility of a qualified licensed guidelines to professional practice in this professional W.neer, or firm of engineers, can and regard."

must be iequired for all steel structures. That respomibility should include design, shop drawing 2.0 DEFINITION OF TERMS review and the observation 01 construction of all The meanings of some terms used in structural sy stems included in a project.

engineering practice and the structural steel industry i 1.2 Background differ somewhat from ordinary' dictionary The system for buildine' steel structures in the definitions. Derefore the following defimtions are United States is complex, involving many parties P ovided in order that certain critical terms used in and vanous contractual arrangements. The Owner. this repon will be uniformly understood.

Engineer. Architect. General Contractor. Fabncator, Design 1.

Detailer. Erector and inspector all perform specific

( As a transitise verb) ne process of utilizing tasks in the design' construction process. Defimng the pnnciples of structural rnechanics and materials

! by contract the scope of work and responsibihties science to determine the size. shape. composition, of the various parties is paramount to achieving

! and arrangement of stmetural components to create i quality and integrity in steel structures. # I" "

in the early part of this centurv there were

~ n nned h m under heet supeMsion of a I

many inconsistent practices for steel construction.

~

9u lified licensed Professional engineer.

In 1924 the American Institute of Steel Construction ( As a noun)De pmduct of he design pmcess.

! ( AISC) initiated its Code of stancaro Practice w hil. as expressed by drawings. specifications, notes or I provided some uniformity. Howeser b) the late other descriptise material.

1960's. changes in technologv and society as a

2. Design Drawings whole began to hase an adverse effect on these Graphic diagrams and details that desenbe the practices. Coon decisions relating to liability and restraint of trade fast-track construction, a hight) proposed v.ructure with sufficient information so that shop drawings can be prepared.

competitise market, high inflation and interest rates.

changes in engineering college curncula, escalating 3. Specifications labor costs. discontinuance of standard connections Written technical requirements which by AISC, and the development of computer Supplement the desiFn drawings and define techniques allowing the creation of more materials, workmanship. quality control sophisticated structural systems all had their effect.

" pr cedures, and identify the gmerning codes /

The resulting new and often confusing trends in standards.

design and construction practices are affecting 4. Contract Documents design responsibilities and led the Professional The design drawings, specifications. Feneral Practice Division of ASCE to appoint the Task and supplementary provisions, addenda, anJ any Committee on Design Responsibility and initiate change orders that define the complete scope of the this study.

Project and the tenns and conditions fa perfammg 1.3 Charge to the Task Committee on Design the contract.

Responsibility

5. Shop Drawings The Charge to the Task Committee is stated Graphic diagrams of structural members and in the minutes of the meetint of the Professional components that show construction informanon and Practice Division Executivt Committee (PRODEX) i, 3

i l

I

serve as the basis for fabrication of the structural 14. Detail 5'##I'

6. Erection Plans ,. I As a transitim wrb) The procen of utilizing the pnnciples of geometry and the an of graphics Assembly diagrams that desenbe the sequence to desclop the exact dimensioning of structural and method of site assembly of the fabricated components so that they can be manufactured by structural components, the Fabricator in accordance with the design

7. Ow ner drawings. Detailing is often performed by non- -

engineers.

The public body or authority corporation.

association. firm or person l'or w hom the structure I As a noun) The product of the detailing is designed and built procen, usually shown on contracts and shop drawings. such as the graphic depiction of a

8. Engineer of Record (EOR) connection.

The licensed professional engineer who 15. Connections deselops the design criteria and framing concept The structural components that transfer forces for the structure. performs the analpis. and is from one member to another within a structure.

responsible for the preparation of the design These include strengthening and reinforcing draw ings and specifications. The EOR is commonly elements for the structural membert identified by the professional engineer's seal on the

16. Responsibilits design drawings. If the ERO n a consultant. he/

she may be either the prime professional. or a Accountability for properiv esecuting the professional associate working for an architect. sers ces and'or for furnishing and crectmp the planner, or construction manager. materiah which the pam is authorued, bs contract.

to perform As used in this report. responubihty

9. General Contractor must cc. esist with authonty-The person. firm or corporation with whom 17. Authority the Owner has entered into an agreement to The pa.ver. c o.*rred or imphed by contract.

construct the project.

to escreise effective direcimn and control over an

10. Fabricator actnity for which a pam has responubihty As used The company that fabricates structural steel in thn report. authority must co esist with responsibihit components in accordance with the requirements of the shop drawings.

11. Detailer Resiew and approsal of tor other appropriate action m respect ofi Shop Drawings. sampics and The individual or organization that prepares other data which the General Contractor is required the shop drawings and erection plans for the project. to submit. but only for cor.formance with the design The Detailer may be the draf ting depanment of the requirements of the structure and comphance with Fabricator or an independent firm to whom the the information gnen in the contract documents.

Fabricator sub-contracts the work. Such resiew s and approvals or other action shall not l 12. Erector extend to means. methods, techniques. sequences The company that erects the fabricated or procedures of construction or to safety precautmns and programs incident thereto.

structural steel components at the site

13. Inspector 3.0 RECOMM ENDED GUIDELINES The organization, firm or testing laboratory The definition by contract of work auign-retained by the owner to serify that the steel ments and responsibilities are prime considera-construction is in compliance with the contract tions for prosidmp integrity in fabricated steel documents.

t structures The Ow ner. Engmeer. Architect. Gen.

l 4

,e, ~ - .-

. . . - - . -_ . - . -. , . . - . , . - . - - - . --- A

eral Contractor. Fabricator. Detailer. Erector. and shop drawings in order to produce safe Inspector are all deeply involved with the integrity ,.

structures.

of the proposed structure. (6) The design drawings should provide The contractual arrangement for design that sufficient information for the Fabri-offers the best control of structural integrity is one cahir to produce correct shop drawings.

under which the EOR has re.sponsibility and au- (7) The EOR should review and approve thority for the entire structural design. including shop drawings prepared by the Fabri-the connections. The guidelines proposed herein cator for comphance w ith the strength should scrse to ehminate confusion and misunder- and stiffness requirements of the standing in the performance of structural steel design.

contracts.

(8) In those cases where the Fabricator requests permission to resisc certain 3.1 Contractual Arrangements connections to facilitate fabrication iIi The EOR should hase responsibilny and or erection, the EOR should re-and authority for all aspects of the siew and approve the Fabricator's re-uructural design. The EOR's contract visions. The Fabricator should should specify that hecshe either Ial reimburse the EOR for costs incurred design the connections. or ibt res iew in makine this resiew and apprme the connections selected 19) Requircrrents that the Fabricator be and detailed by the Fabricator. cenified to perform the work should (2p The Fabricator should hase responsi- be specified by the EOR in the con-bility and authority for properly im- tract documents.

plementing the design drawings, properly furnishing materials and i10) Where the design insohes a non self-workmanship maintaining the speci. supporting steel frame, the contract fied f6nca:ien aM erection te!cr documents should indicate which pany ances. and for fit and erectibihty of the is responsible for the design of the structure. The Fabricator should pre. construction bracing and how king such pare shop drawings in accordance w ith bracing needs to remain in place.

the design information supphed in the (ll) Where crection procedures require contract documents and subsequent special design and calculations, the instrue: ions from the EOR. contract documents should specify that

0) The contract documents should spec. the Erector hase a hcensed profes.

ify that the Fabricator either tal dhtail sional engineer perform these ser-the connections as designed by the sices. The EOR's review of this work EOR. or (b) select and detail the con. should be limited to its effect on the nections for review and approval by the integrity of the main structure.

EOR. (12) On-site observation by the EOR as well

( 8) Fur complex steel structures, the EOR as the services of an independent may specify in the contract docu. agency for shop and field inspection ments that the Fabricator have a li- are aho essential project elements with censed professional engineer design a co31. The Engineer and the Archi-the connections. In such cases, the tect should apprise the Owner of these EOR should still review and approse facts at the time of entering into the the connections. contract for professional services.

(5) The EOR should have sufficient time 3.2 Rationale and compensation to prepare design These recommended contractual Fuidelines drawings and to review and approve should proside the best assurance of quahty and 5

i . . . .

integrity of structural design through the clear as- desiFn drawings or specifications.

signment of responsibilities and authority among

• 6)

( A Fabricator and/or Erector who re-the panies involved with the project. Having the tains a licensed professional engineer EOR involved in all aspects of the structural de- for design of connections and/or erec-sign process will enhance the structural integrity rien procedures should authorize that of the facility and promote the successful comple- engineer to communicate directly w ith tion of the project by centralizing control of the the EOR. but the Fabricator and/or design and decision making process, and provid- Erector should coordinate the work and ing for continuity. keep the General Contractor informed.

On the other hand, contractual arrangements which fail to inc!ude clear assignment of respon- 4.2 Qua:ity Assurance and Quality Control sibihty and authonty for design of structural ele- (1) Contractual relationships between the ments and connections or which tend to impo e pan es involved in the project are vi-on one pany the proper msponsibilities of another talh imponant to its quality and struc-pany are considered to be counterproductive and tural integrity and should follow the should be avoided. concepts outlined in this repon.

(2) Clearly delineated and efficient office 4.0 OTHER RECOMMENDATIONS pocedures are entical to the adminis-tr ti n f design, control and checking 4.1 Communication of shop drawings; and for the process-(1) The contract documents should clearly inc' of design changes: and should be define the lines of communication y;gorously pursued.

among the Ow ner. EOR. General Con- 13) The Fabncator/ Erector should provide tractor. Fabncator. Detailer. Erector, and ongoing inspection and testing to de-I"'P#CI termine that quality is in conformance (2) One or more prebid and/or precon- with the contract documents. This struction conferences should be held, should be augmented with inspection during which lines of communication bv the Owner's inspector and on site and responsibilities are identified, and Eservation by the EOR where ap-the contract requirements are discussed. propriate.

(3) After award of the contract. the Fabri- N) Des pn and construction should be in cator and Detailer should meet with the compliance with all Foveming build-EOR to review design requirements and ing. safety and technical codes / stand-criteria for detailing the connections. In ards to assure quality and structural those cases where the Fabricator plans integrity.

to submit design changes. prior ap-proval should be secured from the EOR. Pirnsrrdin iAc Tad C<wnmmer em keeping the General Contractor  ; d["

mformed. j u g_ ourte, Geerhard Haager (4) The Fabricator and'or the Detailer should **"*"

have a direct channel of communica- ['

tion with the EOR as the project goes !wl R. Munger. PROCEX Contact forward and should keep the General Rmen C. West. PRODEX Conixt C,rresf==Jmg Members Contractor informed.

""H B""h (5) The Fabricator should request clarifi- Stanley D. Lindsey cation in writing from the EOR on spe- waher P Moore. Jr Robert P. Stupp I cial connections or unusual structural conditions not clearly defined by the R Lawrence Whipple. Staff Contac 6

(

s .ea ,

i AbfERICAN SOCIETY OF CIVIL ENGINEERS POLICY STATEAfEAT '

AUTHORITY AND RESPONSIBILITY FOR DESIGN OF STEEI STRUCTURES Approved by the Professional Activities Committee on July 9.1985.

Appro.rd by the hmmittee on Iblicy Review on September 4.1985.

Adopted by tr.e Board of Direction on Omber 20.1985 1

Policy by centralizing control of the design decision-The American Society of Civil Engineers makmg process, and by prosidmp contmuity. Clear belieses that the safety of the public can best be defmition of authority and responsibility by con-protected by having a licensed and qualified en. tract prosides the best assurance for that central gineer perform the design and resiew, apprme the control and continuity. Contractual arrangements shop drawings, and provide appropriate on-site that fail to include clear awignment of responsi-observation of construction of all steel structures. bility and authonty for the design of structural cle-Clear definition by contract of work assignments, ments and connections which diside and fragment authorities and responsibilities during the design such responsibility and atrhonty. or w hich tend to and construction procen is a primary considera a impose on one pany the proper responsibilities of tioa for providing integrity of steel structures. The another party. are considered to be counterpro-owner. engineer of record, architect. general con. ductis c and should be discontinued. The suggested tractor, fabncator. detailer, erector and inspector alg authorities and responubilities are:

have important roles in the design and construc-I. The engineer of record should have re-tion process. and each must do his/her part to sponubihty and authority for all aspects of the achieve this necewary integrity. The contract ar' structural design and hivher contract should pro-rangement for desien that offers the best control of vide suf.kicnt time ar.d o.mpensation for imple-structural integnty is one under which the enpi-menting this role. In addition to design of the neer of record has responubility and authority for structural members, histher contract should spec-the structural design, including the connections.

ify that hershe either (al designs the' connections.

Issue or (b) res iew s and approves all shop draw ings. in-The engineering profession view s with great ciuding connections selected and detailed by the concern failures of steel structures, and ASCE has fabricator. During construction, the engineer of reviewed the engineer's role in design and con. record should make appropriate on site ob-struction practices as related to his/her profes. servations.

sional responsibilities to society. ASCE believes 2. The fabricator should have responsibil-that failures hase resulted. in substantial pan be- ity and authonty for properly implementing the cause of loss of continuity and commumcation as design drawings. properly furnishing materials and a project moved through the phases of design. re- workmanship, maintaining the specified fabrica-s iew of shop draw ings and construction. Esery ef- tion and erection tolerances, and for fit and erect-fort must be made to avoid failures by careful ability of the structure. The fabricator should definition of authority and responsibility by all of prepare shop drawings in accordance with the de-the parties involved in a project. sign information supplied in the contract decu-ment . .,rf subsequent instructions from the engineer Rationale of record. The construction contract documents involving the engineer of record in all as- should specify that the fabricator either (a) details pects of the structural design process should en- connections as designed by the enginecr of record.

hance the structural integrity of the facility and or (b) selects and details connections for review and promote the successful completion of the project anproval by the engineer of record.

7 I /

7 e J